422 
Proceedings of the Royal Society 
under 35° F., and varied from 32° in contact with the ice to 34°*5 
at the bottom. In the long clear nights of a severe winter the 
temperature of the atmosphere may often be 30° below that of 
the surface of the water, so that both by radiation and convection 
cooling goes on with great rapidity, notwithstanding the fact that 
water expands in being cooled below 39° F. ; and it is remarked 
by people dwelling by the side of the lake that the moment ice 
begins to freeze it spreads with great rapidity, so that the whole 
basin is usually frozen over in a night, and as soon as a skin of 
ice has been formed it very rapidly becomes thick enough to bear. 
Epochs of Maximum Temperature . — From the curves and their 
intersections we see that the maximum temperature of the surface 
water occurred before the 5th September. At 13’5 fathoms it 
occurred some time between the 5th and 22nd September, and as 
cooling goes on faster than heating, the date of maximum tempera- 
ture at this depth will be nearer the 22nd than the 5th September, 
probably about the 15th. Similarly at 16 fathoms it occurs 
between the 5th September and 15th October, and probably about 
30th September. At 19 ’5 fathoms the maximum temperature 
occurs about 6th October, and in the second week of October (in 
1885) all water from 20 fathoms to the bottom attained and 
passed its maximum temperature. The maximum temperature of 
the surface occurs about the middle of August, and that of the 
bottom in 35 fathoms about the middle of October. 
Inver snaid Station . — The observations at this station at the 
different dates have, in many respects, greater interest than those 
made in the Luss basin. The depth here is 100 fathoms, and is so 
great that for a thickness of 40 or 50 fathoms above the bottom the 
change of temperature during the course of the season amounts only 
to a fraction of a degree, and is so slight as to elude detection, 
except by using very delicate thermometers. The observations 
made at Inversnaid are collected in Table X., and the results 
are expressed graphically in the curves (fig. 5). On 18th August 
no observations were made exactly on the Inversnaid station, but 
observations were made at two neighbouring stations, Rob Roy’s 
Cave, about a mile north, and Culness, about a mile south of it. 
The curve has been drawn from the means of the temperatures 
observed at these two stations. 
